Rationality with respect to people, places, and times

Abstract

There is a rich tradition within game theory, decision theory, economics, and philosophy correlating practical rationality with impartiality, and spatial and temporal neutrality. I argue that in some cases we should give priority to people over both times and places, and to times over places. I also show how three plausible dominance principles regarding people, places, and times conflict, so that we cannot accept all three. However, I argue that there are some cases where we should give priority to times over people, suggesting that there is impersonal value to the distribution of high quality life over different times.

Keywords:

• Practical rationality
• impersonal value
• decision theory
• temporal neutrality
• impartiality
• dominance principles

Notes

1. Throughout this article, I sometimes speak in terms of place, and sometimes speak in terms of space, depending on which sounds better linguistically in the particular context of usage. But I am using these terms interchangeably, understanding any particular place to correspond to a spatial location.

2. For the sake of simplicity, I follow Parfit (1984) in my presentation, by putting my discussion in terms of being neutral with respect to different ‘people.’ However, in this context, the scope of the word ‘people’ needn’t refer only, or to all, human beings; it might refer to all rational beings or, as it does for many, to all sentient beings.

3. To be clear, and fair to Parfit, nothing in Parfit’s argument against the self-interest theory commits him to denying this claim. To the contrary, Parfit could, and would, accept that there are numerous respects in which we should treat persons, places, and times differently, and he could, in fact, accept many of the claims and arguments that I shall be making in the following sections. The point is just that while Parfit has illuminated an important truth, not to be lost sight of, that in certain respects we should treat persons, places, and times the same, I hope to illuminate a different, but compatible, important truth, also not to be lost sight of, that in certain other respects, we should, and must, treat persons, places, and times differently. In addition, I hope to illustrate some of the sometimes surprising respects in which this is so.

I might add that while Parfit could accept most, and perhaps all, of what follows, Sidgwick, perhaps, could not. This depends on whether or not one interprets Sidgwick to be a mental state theorist, who believes that the only sources of intrinsic value and disvalue are positive and negative conscious states, respectively. That view, combined with Sidgwick’s additive-aggregationist maximizing approach, which assesses the value of an outcome solely as a simple additive function of the individual instances of intrinsic value and disvalue which obtain in that outcome, implies that one must be strictly neutral between people, places, and times, for the purposes of moral reasoning. Such a view is incompatible with most of the claims that I make in the following sections.

However, it is worth noting that it is not the maximizing structure of consequentialism, per se, that is incompatible with much of what follows, but rather the maximizing structure of consequentialism when it is combined with a particular conception of intrinsic value like that of the mental state theory (Here, I focus on the part of Sidgwick’s view according to which there was always sufficient reason, and hence it was always practically rational, to act morally. In fact, as Parfit (2011, 6–7, 130–149) has pointed out, Sidgwick (1907) accepted the dualism of practical reasoning, according to which there was also always sufficient reason, and hence it was also always practically rational, to act self-interestedly.).

I am grateful to Shelly Kagan (personal communication, October 2, 2015) for suggesting that I consider the theoretical underpinnings that might lead some people to think that we should, in fact, be strictly neutral between persons, places, and times for the purposes of practical reasoning. What the preceding discussion reveals is that the positions I shall argue for in the following sections are incompatible with mental state versions of maximizing consequentialism, or other theories of that ilk.

4. I realize that some of these assumptions will be controversial on certain interpretations of modern physics. For example, some believe that one cannot meaningfully distinguish between space and time, as the universe is composed of (inseparable) space/time points. Likewise, some believe that space and time had a beginning, perhaps at the moment of the Big Bang, and likewise, that space and time may have an end, depending on how much matter there is in the universe, and whether the Universe will eventually collapse on itself and everything, including space and time, will come to an end at a single point of singularity. In addition, some believe that time’s passage is an illusion, and others, appealing to Einstein’s theory of special relativity, will insist that the direction of time is relative to one’s point of reference.

Still, there are a number of distinguished philosophers of science, metaphysicians, and physicists who would accept that the assumptions I am making are compatible with our best scientific views of the universe, and it is hard to deny that the assumptions I am making might have been true of our universe or some other universe (I am grateful to my colleague, the philosopher of physics Barry Loewer, who confirmed in an email on 16 October 2015, that the philosopher of physics, Tim Maudlin, ‘thinks that time has an intrinsic direction, and he is willing to say that “time passes,”’ and that some ‘metaphysicians think that time “moves” in a more robust and non-metaphorical sense … [including] Dean Zimmerman … CD Broad, Michael Tooley, [and] Tim Williamson.’ Loewer also noted that ‘two physicists that are in the “time really moves” camp are Lee Smolin and George Ellis.’).

So, one way of interpreting the following arguments is that they may give us pause for treating space and time the same even on our best current scientific understanding of the universe, and would give us reason for treating space and time differently in any universe where something like a Newtonian conception of space and time was true. Accordingly, we should be wary of any a priori arguments in support of the claim that rationality requires that we treat space and time the same for the purposes of practical reasoning.

5. Scheffler (2013) has argued that having descendants who will help realize some of our deepest hopes, projects, or ideals helps to give our lives value and meaning that they otherwise would lack. Scheffler’s views are entirely compatible with my own, and I am happy to accept them. But they point to other reasons why one might be more concerned about the future than about what happens elsewhere in space than those I am trying to illuminate here. As my example makes plain, I believe that even if the future civilizations were unrelated to our own, and would do nothing to further our particular hopes, projects, and ideals, I still believe that there would be strong reason to ensure that such civilizations would exist if they would have high-quality lives. In addition, I believe that such reasons would be stronger than any we would have to ensure, were it possible, that such civilizations obtain elsewhere in space contemporaneous with our own.

Similarly, Jeff McMahan (personal communication, October 2, 2015) suggested a variety of considerations that might lead us, in general, to give greater weight to there being high-quality sentient lives existing in the future, than to there being high-quality sentient lives existing elsewhere in space. According to McMahan, these might include views we have about the importance of the preservation of value, views about the importance of progress, and views about the importance of greater diversity of experiences. My response to McMahan is threefold.

First, as with what I said about Scheffler’s view, I don’t regard my position as incompatible with McMahan’s. Depending on the details of the case, there could be more than one reason for valuing the existence of future civilizations over the existence of contemporaneous civilizations elsewhere in space. But second, in my examples, I wasn’t, in fact, assuming that there was greater diversity of experiences over time than across space, nor was I assuming that there would be progress between our current civilization and the future, unrelated, civilizations. Thus, my views about such cases weren’t, in fact, turning on such factors. Moreover, importantly, I note that the notions of preservation of value, and progress, have a temporal dimension built in to them, but not a spatial dimension. So, McMahan’s suggestions regarding those factors would, if correct, not be a rival to my own, but rather a further elucidation of some of the reasons why we should treat space and time differently for the purposes of practical reasoning.

Finally, Jonathan Weisberg (Q&A, Belief, Action, and Rationality Over Time Workshop, University of Wisconsin-Madison, September 6, 2015) suggested that our intuitions about such cases might reflect the psychological phenomenon of our ‘not wanting the story to end’; perhaps traceable to our early childhood when our parents read us bedtime stories and we didn’t want the story to end, since when it did, our parents would leave, and we would be left alone in the scary dark! It is always difficult to prove that such a deflationary account plays no role in our intuitions about such cases, but when I think hard about such cases, I don’t believe that the best, or main, explanation of my intuitions about them lies in the sort of deflationary account that Weisberg suggests.

This is further supported by cases that I will consider later, where I have similar views about the greater importance of filling past temporal regions, or gaps between present and future temporal regions, relative to the importance of filling other contemporaneous spatial regions, or ‘empty’ gaps between different regions of space that are occupied by high-quality sentient lives. In such cases, it appears that our propensity to ‘not want the story to end’ would have no explanatory role to play as to why we have the intuitions we do.

6. Thoma (2015) notes that there are two different possible ways of ‘filling’ space with sentient life, and wonders which one I have in mind. On one, we add new sentient beings to locations in space that were previously empty. On the other, we make previously existing sentient beings larger, so that they occupy more space, including some previously unoccupied space.

I had the former notion in mind, but Thoma’s observation points to another issue that may have a bearing on how we should regard space and time for the purposes of practical reasoning. It is clear that my thoughts occupy time – each thought begins at one point in time, ends at another, and spans a given time period. It is much less clear that my thoughts exist in space.

As has long been observed, unlike physical objects that clearly do exist in space, thoughts lack extension, shape, or mass. To be sure, I think of my thought as occurring ‘within my head,’ and my head is a material substance that occupies a given region of space; but does my thought occupy the region of space from ear to ear and from chin to scalp? That sounds odd, and not quite right. Does the thought ‘I am hungry’ occupy more space, as one grows from toddler, to teenager, to adult? That doesn’t seem right, either. Similarly, if I were given extraordinary growth hormones, so that I ballooned up to the size of a planet, I don’t think that my thought ‘I am hungry’ would then be planet-sized, occupying the vast spatial region that my body would then occupy.

If this is right, then we see another way that we have to treat space and time differently, for the purposes of practical reasoning. Regarding contiguous future temporal regions, we could result in a net gain in the total amount of time that was filled with high-quality sentient life in either of two ways; either we could ‘fill’ it with new sentient beings who would have high-quality lives, or we could, in principle, ‘fill’ it by extending our own lifespans. But regarding contiguous spatial regions, it seems that we could result in a net gain in the total amount of space that was filled with high-quality sentient life in only one way, by ‘filling’ it with new sentient beings who would have high-quality lives. Were we to ‘fill’ such contiguous spaces either by moving into them, or by making ourselves larger, it seems that either way there would be no net gain in how much high quality sentient life obtained in the spatial regions in question.

Finally, this discussion makes plain that when I talk of ‘high quality sentient life,’ I am referring to the well-being of sentient beings that is related to consciousness. It also suggests that talk of ‘filling’ a region of space with high-quality sentient life is metaphorical, in a way that talk of ‘filling’ a temporal region with high-quality sentient life need not be. I am grateful to Thoma’s observation and query for leading me to consider and clarify my view of these matters.

7. Thoma (2015, 10) questions whether I can legitimately conclude from my examples that ‘filling space is less important than filling time.’ Thoma contends that such a conclusion is indefensible in the absence of a conversion scale that would tell us what length of time is equivalent to what area in space. But Thoma is skeptical whether such a conversion scale is even intelligible. What would it mean, Thoma wonders, to contend that one acre was equivalent to a given amount of time?

Thoma raises an interesting question. But I’m not sure how much I should be troubled by it. The question isn’t whether, for example, a galaxy in space is equal to 100 billion years of time. Rather, the question concerns the extent, if any, to which adding billions and billions of high-quality lives to an otherwise ‘empty’ galaxy (one devoid of sentient life) would significantly improve the outcome, if there were already billions and billions of high-quality lives being lived elsewhere in the universe at the same time; vs. the extent, if any, to which adding billions and billions of high quality lives to an otherwise ‘empty’ 100 billion years would significantly improve the outcome if it is true that there have been or will be billions and billions of high-quality lives lived at different times (perhaps, to keep the situations analogous, at the very same locations to which the ‘extra’ lives would be added).

I don’t believe that we need an answer to the former question – Thoma’s question – in order to answer the latter questions – which are mine. And I believe that the answers that we get to the latter questions, and others like them, support my claim that ‘it is more important for high quality life to be dispersed across time rather than across space’ – at least if that claim is charitably, and properly, interpreted.

Consider someone who claims that freedom is non-instrumentally valuable; that is, valuable itself, beyond the extent to which it promotes other valuable ideals. Suppose that person also believes that pleasure is non-instrumentally valuable, and so believes that eating ice cream is valuable, at least for her, in virtue of the pleasure that it gives her. Echoing Thoma, one might wonder whether there is a conversion scale between the value of freedom and the value of pleasure. What would it mean, one might wonder, to contend that eating a gallon of ice cream was equivalent to a given amount of freedom? Yet, admitting that the question is interesting, and even that it raises a host of deep and important issues, I still think it could be perfectly intelligible, and even true, to assert that freedom was more valuable than eating ice cream!

Note, if we don’t believe that the value of freedom is lexically prior to the value of pleasure – and I don’t, as, setting health issues aside, I would gladly trade a tiny infringement on my freedom, for a lifetime of ice cream indulgence! – we may need to say a bit more to convey exactly what we have in mind in saying that freedom is more valuable than eating ice cream. But I trust that this could be done. For example, we might just mean that a lifetime devoid of ice cream eating would be better than a lifetime devoid of freedom. But the point is that we might be able to perceive the truth of such a claim, even in the absence of having a conversion scale of the sort to which Thoma is referring.

Similarly, I might not have a conversion scale between areas and times, and yet I might be correct in claiming that ‘it is more important for high quality life to be dispersed across time rather than across space.’ The content of my claim can be derived from the examples I presented in support of it, though admittedly, perhaps only imprecisely. That content is, I believe, intelligible and defensible, and does not require that filling ‘empty’ times with high-quality life has lexical priority over filling ‘empty’ spaces.

There is, of course, much more that could be said in response to Thoma’s worry, but, given my space constraints, I hope to have said enough for this article’s purposes.

8. Indeed, by some standards of measuring how much of each atom is ‘occupied’ by material entities, an atom is 99.9999999999999% empty space. See, http://physics.stackexchange.com/questions/39143/percentage-of-water-that-is-void-or-empty-space.

9. Shelly Kagan (personal communication, October 2, 2015) wonders how far I would push the idea of the importance of filling empty times with high-quality sentient life. Suppose, for example, that there was a planet teeming with high-quality life, such that for 12 hours each day everyone was awake, and experiencing high levels of consciousness, but that for another 12 hours each day, everyone fell into a deep, dreamless, sleep, utterly devoid of conscious experiences. Would I think it a ‘great tragedy’ that there was so much ‘empty’ time over the course of each year? Would the world be better if half of the people had their sleep cycles shifted, so that at every moment, there were always large numbers of people experiencing high levels of consciousness? What if there were only five minutes each day, or five seconds each day, where everyone was completely unconscious? Would the outcome really be better, even in such cases, if people’s sleep cycles were shifted so as to fill every available period of time with high levels of consciousness? Kagan finds this hard to believe.

I confess that I am tempted to the view that, if possible, it would always be better, other things equal, to fill any empty period of time, no matter how short, with many instances of high levels of consciousness; though, obviously, the shorter the period is, the less important it would be. But I’m not committed to this view. What I am committed to is the thought that it is bad if there are significantly lengthy periods of time – leaving it open for now whether we understand that to be on the scale of a decade, century, millennium, or eon – which are utterly devoid of high-quality life. It is better, I think, if there are large numbers of high levels of consciousness obtaining within, and spread throughout, each such period. But whether it is important that there be no gaps at all of high-level consciousness within each period, or exactly how small such gaps would have to be for them to be unobjectionable, are questions about which I have no firm views or commitments.

10. This example is a variation of an infinite one suggested to me by Jeff McMahan (after my lecture at the Oxford Moral Theory Seminar, May 18, 2015) in defense of the view in question. McMahan noted that he thought that a universe where, for all of eternity, there were billions of people in extreme agony, and no one else existed, would be worse than a universe involving an infinite number of planets, each populated by billions of people in extreme agony, and no one else existed, if, in the latter case, all of the universe’s agony obtained within a single short period, such as 30 or 100 years. I agree.

Frances Kamm (personal communication, October 2, 2015) indicated that she might value a universe where lots of conscious life existed for all of eternity, over a universe where an infinite amount of conscious life existed for only 30 or 100 years, even if all of the conscious life was suffering greatly. But when pressed, it became clear that Kamm only thought this on the assumption that, all things considered, there was value to the existence of conscious life, even if that life involved unrelenting suffering. McMahan and I were assuming differently in making our claims. But if we agreed with Kamm, then our cases would be similar to my initial cases, where the argument purported to show that it was more important to fill time than space with conscious beings if, all things considered, the lives of those conscious beings were valuable.

11. There are at least two reasons one might think this rational. First, following Nagel (1986), one might believe that there are different perspectives on the world that one might rationally take, each of which reflects its own reality. In particular, one can take a (more or less) ‘objective perspective,’ which reflects the world as it is, from the ‘outside,’ as it were. In addition, one can take a (more or less) ‘subjective perspective,’ which reflects the world as it is, from the ‘inside,’ as it were. Even if one grants that from a certain objective perspective, there would be nothing to choose between the two alternatives, it seems undeniable that considering the lived life as it would be experienced from the inside, there would be every reason to delay the timing of one’s entrance into Hell as long as possible! On Nagel’s view, both perspectives reflect reality, and both are sources of genuine reasons to believe, care, or act. A fortiori, on Nagel’s view, it would be perfectly rational to delay one’s entry into Hell.

Second, one might grant that there are various perspectives and theories which would regard the life with the delayed entrance into Hell and the life where one enters Hell immediately as equally bad. But surely there are some perspectives and theories that would regard the life with the delayed entry into Hell as better than the life that begins in Hell immediately, while there would appear to be no perspectives or theories that would rank the life that begins in Hell immediately as better than the life with the delayed entry into Hell.

Accordingly, even if we attach most credence in the theories that regard both lives as equally good, as long as we attach some credence in the theories that regard the delayed entry life as best, and no credence in the theories that regard the immediate entry life as best, then from the standpoint of practical rationality, we have everything to gain and nothing to lose by following the theories that favor the delayed life. This follows from simple dominance-type reasoning, since if, in fact, a theory that regards the two theories as equally good is true, then we won’t be any worse off if we pick the delayed entry life, while if, in fact, a theory that regards the delayed life as best is true, we will be better off if we follow it, and worse off if we don’t! A powerful argument in support of the practical rationality of choosing in accordance with the line of reasoning suggested is presented and defended in Ross (2006). Ross’s argument is discussed and employed at numerous points in Temkin (2012, 35–36, 40–41, 125–127, 171–173, 261–262, 443–445).

12. Note, our previous discussion would support a similar, though possibly weaker, conclusion for finite cases. If it is more important to fill temporal locations that are devoid of high-quality sentient life than spatial locations that are devoid of such life, it would be more important to shift the temporal locations of some people in an overcrowded world to some point in the future that would otherwise be devoid of high-quality life (say, via suspended animation, if that were possible), than it would be to merely shift the spatial location of the people in question by sending them to another planet that would otherwise be devoid of such life (say, via teletransportation, if that were possible). For further discussion of this position, see Section 6.

13. The following case is my own, but it was sparked by an example I first heard from John Broome, many years ago, which he called ‘Expanding Heaven and Expanding Hell.’ Broome credited his example to Cain (1995). Although my views about this topic were arrived at independently, other philosophers have developed similar arguments in order to make similar points. See, for example, Vallentyne (1993), Cain (1995), Lauwers (1997), Vallentyne and Kagan (1997), Machina (2000), Lauwers and Vallentyne (2004), Bostrom (2011), and Campbell (2015).

Interestingly, while Cain (1995) uses an example similar to mine to arrive at the same conclusion that I do regarding the relative status of Personal and Temporal Dominance Principles for certain cases and contexts, Campbell (2015) produces a series of ingenious examples in order to show that, depending on one’s theory of personal identity, there will be other cases where the relative status of Personal and Temporal Dominance Principles would be the reverse of what Cain and I argue for. I don’t happen to favor the reductionist view of personal identity that would lead to Campbell’s results, but many do, and for those who do, Campbell’s arguments are quite compelling.

14. As recognized in note 13, Campbell (2015) has shown that, depending on one’s view of personal identity, there may be cases where the judgment yielded by the Personal Dominance Principle seems false. Moreover, in Section 6, I will present other cases where the Personal Dominance Principle seems false, that don’t depend on one’s views about personal identity. Thus, on reflection, I believe that the Personal Dominance Principle needs to be revised, or limited in scope, and the same is true of the other dominance principles.

15. For reasons of the sort adduced in note 11, assuming that we give some credence to thinking that the Personal Dominance Principle applies in such a case, then for the purposes of practical reasoning, we should choose as if that is the correct theory, even if in fact we give more credence, and even much more, to the Impersonal Neutralist View. This is because, for Pareto-like reasons, in this kind of case, we have everything to gain and nothing to lose by following the recommendation of the Personal Dominance Principle. See Ross (2006), and Temkin (2012, 35–36, 40–41, 125–127, 171–173, 261–262, 443–445.

16. My use of the terms ‘narrow person-affecting view’ and ‘wide person-affecting view’ vary in certain important respects from Parfit’s use of those terms, but in ways that need not concern us here. See Parfit (1984, 393–401; Temkin 416–45).

17. I am grateful to Frances Kamm (personal communication, October 2, 2015), for leading me to see that there were a slew of issues of this sort that ultimately need to be considered and resolved.

18. The methodological approach of seeking ‘reflective equilibrium,’ famously championed by Rawls, was, as Rawls himself acknowledges, previously employed by Sidgwick (1907).